To think ChatGPT is the best thing ever?

I genuinely have no idea what I’d use it for.
SIL uses it to write school reports.

Don’t be silly. As long as people are aware that it exists, what it does, and how to use it, they are not being a dinosaur by making an informed decision not to use it. I agree that people shouldn’t stick their heads in the sand about the existence of something, but nor should they blindly use something without question.

When I first heard of it I thought it was amazing. But having heard so much now about the negative impact on the environment from AI, it really worries me. I also think it encourages people to be lazy and to never think for themselves.

Most of that you could have found via an internet search. I think what chat gpt and others do is basically stop us from using our own brains to solve problems. This might be a great relief initially but over time the societal consequences are scary especially as it often lies or hallucinates

Totally agree with you! I used it last week to draft a really awkward email to a company that had been messing us around for weeks. I just typed "make me sound polite but firmly annoyed" and it nailed it in about two seconds flat 😂 Saved me ages of staring at a blank screen stressing over the wording.

But I am completely with you on the AI pictures. Those weird fake photos on Facebook of kids building giant things out of plastic bottles give me the absolute creeps! It's brilliant for admin though.

Assumption is the mother of all fuck ups, as Apple, Google and Microsoft can't even make their own products work correctly and securely, why do you think their checking of yours is any better. I've lost count of security upgrades to windows and MSOffice, there seems to be an IOS update every week now.

It is useful to check your thinking where you have done the research and remain in control. But it can’t tell where it has gone wrong or failed to account. It is best when already clear guidelines exist.

I am currently using it to check my position statement for FDR in divorce and it’s confident with section 25 and how that applies in a needs based case because it’s all over the web. But for individual stuff it can get things wrong in a way that’s not immediately obvious but strategically significant and when questioned can be strangely defensive suggesting that MY previous calculations were incorrect when referring to its own calculations. Or it flatters me that I made a good catch. If I took its calculations and therefore strategic positions as gospel I would look like an idiot. I basically have to fact check everything and perform my own calculations independently to make sure it’s not giving me a bum steer.

But still preferable to having no sounding board or paying hundreds of pounds to a solicitor for generic advice. I couldn’t afford to pay a solicitor to calculate the advantage/disadvantage to both parties of a higher pension share versus equalising income on retirement share plus maintenance in a long marriage with DB pension and no other assets where one party was unable to work or accrue further pension due to caring responsibilities.

I’m not convinced that AI is an advantage to the majority of the increasing amount of people that can’t afford legal representation and don’t qualify for legal aid.

I would assume that the checking and acceptance process by Apple/Google would have identified any such huge problems and refused to make the apps live tbph. They didn't, however.

In terms of the app working properly - it does, so far.

It's not glorified Google. The way it works is quite different and that difference is key. unlike Google it's not a keyword search at all. In fact it's not a search at all. It generates text by always trying to predict the next word but this ability is based on text it has been trained on which is why people are confused about its difference to search engines. When you send a query to chat gpt it doesn't just bring up text it finds on some website but it generates text. This means it is really good at text generation obviously but not so great at search, which might be why a simple Google search is more useful when you just want to look for existing text.

And yeah data centres used to train language models are terrible for the environment.

I'm still amazed at how well it works.

So you have no idea if the code works properly and traps errors, what it does with user data or if it is riddled with trojans, viruses and ransomware. What is your plan when all the devices with it on freeze and all users are locked out, with their bank accounts being emptied.

I used it to calculate the interest rate on something (as I couldn't be bothered to do it). It was good for that, but I work in the creative industry and it's terrible at coming up with writing that works in the correct context. It doesn't 'think', understand context or have critical thinking, so it just tells you what you want to hear. It also uses up a huge amount of electricity and water

Nope. None...I mean literally zero. We copied and pasted huge reams of code hundreds of times, when and where told to, and when the page wouldn't save and threw up an error we screenshot it and told Chat GPT to fix it.

It's easy to avoid. Forever, the first results on Google have been the ones that pay to be there and are rarely the ones you want because they'll be crap in some form. So it's instinctive the scroll past until you see the proper website or other reliable source.

Do you mean that Google AI blah which you have to trawl halfway down the page to ignore, it is very often wrong. I always have to remember now to avoid it.

I intend to keep all of my faculties into old age (a long way away). Use them or lose them, if you're outsourcing your thinking to AI then you'll end up unable to come up with an original thought.

People who learn things for themselves and don't outsource everything on life to their phone will never be in the dark ages.

I prefer CoPilot to be honest. Although CoPilot has too many morals and won’t do some of the things I ask it to for work but ChatGPT will happily do them so I use it then. 😆

im just about to start an apprenticeship in AI. Looking forward to it.

If you ask it not to make things up in settings,

It astonishes that this would even be necessary. Surely (unless you specify that you're looking for a work of fiction) it shouldn't be making things up in its answers.

Ok fair enough. You are already a dinosaur.

Yawn.

You are going to be left in the dark ages. It’s everywhere. You may not like it, but if you don’t start embracing it, you will be considered a dinosaur.

It’s ok, I wouldn’t go as far as saying amazing yet. It’s way too much of a sycophant and it can’t count to 200 which is weird. I saw a YouTuber trying it so did it myself and it just can’t do it. So yeah, it’s not ‘amazing’ but it’s useful at times.

It takes 10,000 litres of water to make a pair of jeans and fast fashion has killed many lands. It’s the 2026 version. We couldn’t beat fast fashion but has likely passed its peak

@Daisychain01 yes, I do realise that. What I was saying is that I prefer to see the full results of a search (often over multiple pages) rather than rely on an AI summary of the most popular results, many of which are often tripe.
I hope you didn't mean to sound quite as patronising as you did.

The water needed to cool AI servers is likely to contribute to water shortages across the planet in future. According to Google, 500 milliliters (about 16 oz) of water are used for every 5–50 ChatGPT prompts